Thank you for you input!
In that comment of his, he referenced an old post in which he compared his WTA to a 1 million sinc filter. Was that filter suppose to resemble your Sinc-M?
Anyway, I can understand that he sees HQPLayer as a competing solution, but I really don’t appreciate his bluntness when he calls it wrong without presenting evidence to that.
He makes $$ for every Chord DAC sold.
He has a very large and loyal customer base.
The larger his loyal following grows, the more $$ in his pocket.
So of course he will say his DAC design and digital filtering solution is the best ever and nothing else is good - this also results in more $$ in his pocket.
I know. That’s why I confrunted him.
So I just downloaded the trial of HQPlayer and got it playing with Roon after tinkering.
Can someone take a look at my settings and let me know if these are the best settings to use that would be closest to the M Scaler with my TT2? Any suggestions would be appreciated:
Add images
sinc-Ll is technically closest to M Scaler as filter type. But you have better choices as well, such as the sinc-M you have selected. Or even better for example poly-sinc-gauss-xla. So also give other filters a try, you can use the table in manual to help choosing potential candidates.
Thanks. Will give it a try.
I’m interested in trying out HQP and am interested in your advice about how to change my playback chain to do that. Currently this is Roon Core on Synology DS418play > LAN WiFi > SoTM sms-200 with WiFi dongle > USB > M Scaler > Wave Storm dual coax cables > TT2 > HD800S. I’ve got a used Dave arriving in a week to replace the TT2, so I’ll be up for tweaking and testing for a while to make the most of the new arrival.
First I need to introduce something to run HQP. I have a spare Mac mini (Late 2014 i5 2.6GHZ 8GB) that I guess is up to the job of handling upscaling to 768kHz, and DSD512 too subsequently if that’s likely to be even better?
I could simply substitute the Mac mini for the sms-200 for testing, but should I also introduce another spare bit of kit I have, a fanless PC (ACEPC Mini PC Windows 10 Pro Intel Atom x5-Z8350 Processor 4GB) to run the NAA? I understand from a post on another thread that a Window 10 device would be needed to get to DSD 512 from 128, but there may be other pros and cons as well (in addition to adding in another bit of kit that could go wrong!)
Finally, this is some way off, but if I like what I hear, then the M Scaler becomes redundant for upscaling. But simply removing it and the Wave Storm cables to plug directly into the DAC USB would reintroduce the RF noise problems, and maybe some others. I’d like to understand what options might be recommended. Thanks
My 2 cents here (FWIW I daily drive hqplayer with Qutest)
Your Mac Mini is certainly good enough for 768kHz PCM up-sampling (what M scaler does). It’ll probably handle most dsd256, but very little dsd512. But if you are using with Chord TT2, then PCM 768kHz is what you want anyway.
As long as your sms-200 supports NAA (HQPlayer streaming endpoint), then it should be fine. Just need to confirm that it supports NAA. I would recommend against streaming hqplayer via WiFi. It is very high data rate at 768kHz, and can easily glitch on most wifi system.
Regarding connection to TT2, USB is fine for experimenting. But I highly recommend checking this out from audiowise:
It’ll convert USB to dual BNC, and directly connect to the dual BNC on your TT2. I had a noticable improvement with this over a direct USB connection. (Chord USB implementation isn’t very good, it’ll glitch sometimes on 16x PCM and play static).
Beyond the above for signal chain, the hqplayer software setup is pretty straight forward. Be sure to set to 24bit (not the 32 bit default) of you use the dual BNC converter. And then you can go wild having fun with filters and noise shapers. I like sinc-L w/ LNS15. But ymmv.
Adding a bit to the earlier response…
If you use Chord’s internal USB interface, which can be tricky, it may be more reliable to use fixed 705.6k rate. This should make Chord’s trademark white noise blasts less likely. That SRC-DX is a way to work around such issue.
There is no point in sending DSD to Chord DACs, since those will always get converted to 705.6k PCM with less than perfect algorithm. So also for playing DSD sources, let HQPlayer to do the conversion to PCM first.
For Chord DACs, using fixed output rate is particularly beneficial, since the DAC itself always runs out of single fixed frequency clock. It doesn’t have separate clocks for 44.1k and 48k families. Instead, it uses frequency where common multiple for both rate families is only a little bit off.
Thanks for the reply. The sms-200 does support NAA, so your suggestion makes good sense to trial 705/768kHz. My UniFi network reports that the sms-200 is connected by WiFi5 at around 400Mbs, which should be fine to handle 768/24 (which is around 33Mbs I believe). In practice I have no convenient ethernet where it’s located, so I’ll test with WiFi and see how it goes.
I’ll take a look. This sounds like an option that would allow me to remove the M Scaler if HQP sounds as good or better to me. I know the TT2/Dave are reportedly great at rejecting jitter, but I don’t understand how they can deal with long term clock drift if they can’t control the speed of the source, which they can if connected by USB. I guess at the moment the M Scaler clock is the one determining the rate of data coming down from the sms-200.
That’s the plan! Thanks again
I can see this issue has been reported by some, and disputed by others. The sms-200 is linux-based, and I had no problem using it directly connected to the TT2 before I got the M Scaler. Mind you, I wan’t upscaling to 705/768kHz at that point.
I need to read more to understand this better. I simply saw another thread on here talking about this, with some contributors saying they preferred DSD512.
That doesn’t sound good at all. Can you please explain more?
According to Rob Watts, DAVE is the only one that doesn’t decimate DSD when in DSD+ mode. Although he doesn’t explain exactly how he converts this DSD to analogue.
But as he says himself, all his other DACs do convert DSD to PCM705kHz .
Given this limitation with Chord DACs, is 705.6kHz the best fixed rate option to choose?
Out of curiosity, do you know how the M Scaler handles this? Fixed output rate? Or does it up-sample using the rate family of the source file? Nevermind, answered this by reading M Sacler user manual. It up-samples within the rate family.
@jussi_laako (or anytime else who knows):
the Qutest and Hugo2 (and I’m assuming other Chord DACs) have the option to enable a “high frequency filter beyond 20kHz”. This seems very similar to the HQPlayer 20kHz filter option, which makes a lot of sense as a way to mitigate poor HiRes recordings.
When I am playing upsampled 16x PCM from HQPlayer, would you recommend leaving this “20kHz” filter in the Chord DAC turned on or off? For filtering high frequency noise from HiRes source material I am using the 20kHz filter in HQPlayer, so using both seems redundant. Thanks!
The user guide mention it (this is the Qutest user guide)
Measurements are more useful than their words. It looks to be a low pass filter, green filter similar in general concept to HQPlayers 20kHz filter. Although filter design obviously different (steepness, amount of fitlering etc)
Leave it off, it won’t improve anything.
@elementze have you worked out how to turn it off? I’m not sure what the best setting to turn it off is. It seems like by some use of logic when reading the four setting’s blurb in the manual it is best to use Red because it is the only one that is neither WTA1 or WTA2. Or if I’m sending 705/768 kHz is the filter disabled anyway?
Yes, first filter stage is bypassed when you input at such rates. This is also what MScaler does.
